Participation in treatment has consistently been shown to reduce substance use and associated risks, but treatment gains are often not maintained after discharge. "Step-down" aftercare has intuitive appeal and preliminary empirical support; however, existing studies are restricted in generalizability, particularly for drug dependent clients after outpatient services. Twelve-step groups (12SG), the most widely available and least expensive form of aftercare, have been found helpful in extending treatment gains. Longer duration and higher level of 12SG affiliation are associated with better outcomes; however the dropout rate tends to be high, particularly after the first 6-months and some individuals never attend. Specific formal treatment approaches can enhance client's post-treatment 12SG affiliation and attendant benefits but little is known about predictors of 12SG affiliation, about barriers to affiliation, about the process of affiliation or reasons for disengagement. A greater understanding of these issues can enable professionals to focus practices in ways that would facilitate clients' transition into 12SG after treatment. Existing 12SG studies have been limited in sample, scope and methodology; in particular, there have been few studies of illicit drug users (compared with alcoholics); and women and ethnic minorities have been under-represented. Only one study has examined affiliation with Narcotics Anonymous (NA), and self-selection has not been adequately examined in 12SG effectiveness studies. The proposed study is a prospective, process-oriented investigation of predictors of NA and AA affiliation and their effectiveness on substance use and related outcomes after outpatient treatment discharge. Study subjects (N=300) will be randomly selected among new admissions into 2 large publicly-funded inner-city drug treatment programs in New York City. Assessments will be repeated at discharge, and at 3-, 6-, and 12-months post-discharge. The specific aims are: 1. To identify predictors of affiliation with NA and AA among drug users at 3-, 6, and 12 months after treatment. 2 a) To assess the effectiveness of 12SG affiliation on proximal and distal outcomes during the 12-months post treatment period adjusting for self-selection biases; and b) To examine the processes underlying the effect of 12SG affiliation on outcome over time. 3. To gain a greater understanding of choice of fellowship (NA and/or AA), obstacles to affiliation and disengagement from 12SG among illicit drug users by conducting in-depth ethnographic studies. 4. The analyses to address these aims will be guided by an empirically-based conceptual model. Subgroup analyses will be conducted for gender and ethnic subgroups as well as for NA and AA affiliation.